Cities Network Along the Silk Road

Pengfei Ni Marco Kamiya Ruxi Ding •

Cities Network Along the Silk Road The Global Urban Competitiveness Report 2017

123

Chapter 14

The Conjunction of Networked Agglomeration and Location Factors in Chinese Cities: Taking FDI and Domestic Investment as an Example Shuai Shi, Ronald Wall and Kathy Pain

14.1

Introduction

Taylor has contended that according to Jacobs (1969), ‘cities in networks’ have existed for thousands of years (Taylor 2012). Cities can be regarded as the original form of cooperation, agglomeration and international trade in human society and economy. Cluster and agglomeration processes within cities, combined with network/city connectivity processes between cities (Taylor 2012), have improved economic productivity by generating integrated markets, labour pools, new technologies and innovations. There is no doubt that agglomeration and network effects together make contemporary cities the critical inter-connected nodes for human reproduction, creativity and economic growth. Contrary to late twentieth century predictions of the “end of geography” (O’Brien 1992) and “death of distance” due to the rise of advanced telecommunications and linked technological breakthroughs (Cairncross 2001) generating the “network society” (Castells 1996), the United Nations (2013) has predicted that 64.1 and 85.9% of the developing and developed world respectively, will be urbanized by 2050. And as predicted a century ago by Gottman in relation to the United States, most of this population growth will occur in big cities that are spilling over metropolitan boundaries to form extensive globalizing ‘mega-city regions’ consisting of many functionally inter-linked geographically proximate large and small urban settlements (Gottman 1961; Scott 2001; Hall and Pain 2006). Many scholars have dedicated themselves to explaining urban agglomeration phenomena (for example, Marshall 1920; Isard 1956; Helsley and Strange 1990; Krugman 1997; Glaeser 2010). In 1920, emphasizing the importance of specialization, Marshall articulated three sources of agglomeration economies: labour pooling, scale economies of intermediate input and tacit knowledge spillovers. In contrast, Jacobs (1984) has argued that it is local diversification that drives agglomeration externalities, emphasizing as priorities heterogeneity versus homogeneity and parallelism versus hierarchism respectively. Jacobs has asserted that © China Social Sciences Press and Springer Nature Singapore Pte Ltd. 2017 P. Ni et al., Cities Network Along the Silk Road, DOI 10.1007/978-981-10-4834-0_14

235

236

14

The Conjunction of Networked Agglomeration and Location Factor …

knowledge spill-overs generated across diverse industries are more active and innovative than those generated across specialized clusters. Ellison and Glaeser (1999) found that in North America the distribution of diversification and specialization is not random but highly concentrated in big cities with four-digit sectors.1 Meanwhile, it has been found that larger cities tend to be more diversified and specialized in service sectors as opposed to manufacturing whilst many cities also show a significant path dependence tendency and location inertia (Henderson 1991; Kim 1995). This conclusion has been endorsed by research into advanced business (producer) services (APS) clustering in North West Europe where two distinctive urban processes have been identified: Process A—‘mega-city region economic expansion’ and Process B—‘mega-city regions of proximate cities’ (Taylor and Pain 2007; Pain 2008a). Process A is distinguished by functionally polycentric multi-sector clustering which generates Jacobsean economic enveloping and upgrading of towns and cities surrounding a major globally networked city, whereas process B is distinguished by morphological urban polycentricity and sectoral specialization which generate more static, less globally networked, regions (Pain 2008a). Both Jacobs (1969) and Castells (1996) have emphasized city agglomeration as a process that involves flows between cities. Among such flows, investment inputs, throughputs and outputs, are significant in the reproduction of capital, the allocation of resources and so the stimulation of business activities and employment that diffuse technology, knowledge etc. (Domar 1946; Romer 1986; Barro 1989; Anderson 1990). Foreign direct investment (FDI) and domestic investment (DI) are therefore valuable indicators that can shed light on agglomeration economies in urban areas and city regions. FDI is a relatively new form of investment under prevailing globalization, characterized by transnational practices, international exchange, complex ownership, long-tern intentions and so forth. It is also argued that FDI generates technology and information diffusion, elite and skilled-labour flows, complementary capital and export incentivization with international attributes (Noorzoy 1979; De Mello 1999; Kim and Seo 2003). FDI by multinational corporations (MNCs) especially, contributes to job generation, trading and the transfer of capital, skills and technologies (Borensztein et al. 1998; Blomström and Sjöholm 1999; Liu 2008). Mello (1999) has denoted that FDI has positive incentives for domestic firms by upgrading their capital stock. It is estimated that 30% of the productivity growth of manufacturing in the UK was affected by FDI between 1985 and 1995 according to Barrell and Pain (1997). However, the positive effects of FDI are not permanent and some scholars are skeptical as to the compelling importance of FDI flows into domestic markets. Industrial organization theorists have expressed concerns that the MNC strategy of increasing FDI to compete for control of global value chains weakens the economic sovereignty of host countries because the competitiveness of indigenous firms is diminished (Caves 1971; Dunning 1981). Aitken and Harrison have found that fiercer competition introduced 1

The standard industrial classification by 4 digit codes in the US, established in 1937.

14.1

Introduction

237

by foreign firms, crowds out the market share of domestic firms. Hejazi and Pauly’s analysis of the effects of inward FDI in Canada discovered negative impacts of FDI on the domestic market. There is a contradiction then between this skeptical interpretation of FDI and theorization of cities and mega-city regions as an economically vibrant process made possible by flows within global networks of firms and FDI.

14.2

Explaining the Urban Agglomeration Process: Location or Network?

This contradiction is important considering the spatial implications of agglomeration/cluster processes within cities and network/connectivity processes between cities as noted in North West European mega-city regions (Taylor and Pain 2007). Fingleton’s (1999) spatial auto-correlated error model estimating regional economic convergence found technology spillovers across 178 European regions. Van Oort (2007) has estimated the effects of agglomeration economies on neighbouring regions by means of a spatial dependence model which has found significant spatial dependence of the growth externalities of cities. Jocobs et al. (2011) have asserted that the spillover of complementary knowledge across diverse urban industries is the most important source of innovation and agglomeration economies. Their analysis of a sample of 459 cities across the world found that advanced maritime producer services tend to be attracted to the cities where their clients or advanced service providers locate. However, in contrast, Boschma (2004) has argued that knowledge externalities are geographically bounded due to privileged access to information flows, knowledge transfer and interactive learning. And in terms of American firms specifically, Henderson (2003) has found that the spatial implication of agglomeration economies is restricted according to the characteristics of industries: positive technology spillovers for high-tech firms, but not for manufacturing firms. The source of externalities and their spatial implications has become of key interest in relation to agglomerated activities such as FDI and DI with contrasting perspectives presented by the two main theoretical schools explaining the spatial agglomeration of economic activities: location factor analysis (city endogenous growth) and connectivity analysis (the city network paradigm). Diversified location factors are highlighted as endogenous drivers, such as local market size, labour pool, accessibility, industrial configuration, institutional context, high-tech clusters, cultural atmosphere and urban landscape such as universities, urban lifestyles and diversity (Florida 2002). Turok (2004) has conceived the focus of urban development strategy as transferred from spatial policy to the exploitation of indigenous strengths. According to the urban competitiveness literature, location factors are critical in boosting industrial innovation and upgrading capacity which are seen as core competitiveness factors under processes of product differentiation and globalization (Porter 1990). However, over-emphasis on the competitiveness of

238

14


nations and firms gives rise to a missing space between the national level and the firm level. The city level is regarded as a suitable ‘wedge’ to fill this space due to predominant agglomeration effects under pervasive neoliberalism and the intermediate role of cities between the nation and active actors, firms and institutions. Amongst others, Boschma (2004) has asserted that, like firms, cities compete with each other under conditions of strong economic specialization in similar markets. Along the lines of Jacobs’ analysis (1969), it is variety that leads to knowledge creation, and learning has extended from an organizational to a territorial level. Generally, the capacity to attract investment through location factors is one significant aspect to explain economic competitiveness. It is argued that the impacts of FDI depend on the absorptive capability of the domestic markets of host countries (Agosin and Machado 2007; Mahroum et al. 2008). Specifically, human capital, financial markets, and technology gaps are critical location factors explaining the determinants of absorptive capacity (Glass and Saggi 1998; Alfaro et al. 2004; Mahroum et al. 2008). For instance, Blomstrom et al. analysis of the effects of FDI in the UK (1999) has found positive effects in England but negative effects in Wales and Scotland. They have attributed this result to the technology gap between foreign firms and domestic firms. On the other hand, there are those who criticize the epidemic of urban attribute (location factors) analysis and econometric studies in explaining urban development (Berry 1964). They deem that markets stem from social networks intrinsically whilst market regulation reflects hints, trust and rules generated by mutual communication of producers’ and marketers’ networks. It is argued that the network is an invisible space where ideas, thoughts, innovations and learning can be generated and shared (Powell et al. 1996). Nevertheless, this theoretical debate does not mean that the two analytical approaches are mutually exclusive. In fact, they can be seen as complementary since networks underline connectivity whilst location factors provide endogenous thinking. Berry (1964) has claimed that urban geography essentially focuses on ‘cities as systems within systems of cities’, which indicates that we need to look at cities in a broad perspective emphasizing their connectivity, such as investment relationships between cities. His view is broadly in line with the emphasis on inter-city relations of Jacobs (1969). However, the networks of cities are spiky; cities hold different centralities according to headquarter-subsidiary structure and distinctive functions (Wall and Van der Knaap 2011). Moreover, due to fuzzy production modes, different specializations, and the division of labour, the economic and functional relationships between cities tend to be more complementary than competitive, benefitting from scale economies, knowledge exchange and synergies (Capello 2000; Meijers 2007). Especially under globalization, big cities are more globally connected and have more relations with other big cities than with their neighbouring cities (Pain 2008b; Pain and Van Hamme 2014). Johansson and Quigley (2004) have agreed that agglomeration and networks are complementary to each other in terms of knowledge diffusion and productivity gains. Nevertheless, we cannot ignore the links between knowledge diffusion and investment because investment, especially FDI, is regarded as an important pipeline to diffuse knowledge. Therefore, it is necessary to observe the agglomeration of

14.2

Explaining the Urban Agglomeration Process: Location or Network?

239

urban activities such as the agglomeration of FDI in a network way. Accordingly, this paper tests the pipeline function of networks in FDI agglomeration. In addition, the dispersal of production modes and the concentration of specialized services is evidence that location factor analysis and network analysis cannot be separated when we analyze the economic activities in cities, especially capital flows. Burt (2009) has highlighted that competition is a relationship issue other than the competition of economic players themselves. Specifically, it is argued that establishing linkages and resultant collaboration networks can help firms to obtain access to external knowledge and boost regional internal productivity as an outcome (Powell et al. 1996; Nicolini et al. 2003). Cooperation with external partners is tremendously meaningful for capability in product development, particularly in business markets (Håkansson and Snehota 1989; Pain 2008b). The introduction of social network analysis (SNA) complements the traditional focus on attributes and location factors; it emphasizes the relationships among actors instead of the actors themselves, and aims to incorporate topology to illustrate the structure of networks (Borgatti and Foster 2003). Accompanying accelerating urbanization and integration in global markets, China has become both the second biggest economy and the biggest FDI host country in the world. Through a time series analysis, Tang et al. (2008) have asserted that FDI has a positive effect on Chinese economic growth and the domestic market, though DI has less impact on FDI. In addition, Ouyang and Fu (2012) have noted the inter-regional spillover of FDI from the highly urbanized eastern coastal areas of China to inland areas, meanwhile the capacity of inland cities to absorb spillover profits is mainly based on their manufacturing and mining capacity. This paper does not only follow a series of previous studies of location factors and ‘crowd-in’ or ‘crowd-out’ of FDI and DI, but it attempts to explain the agglomeration of FDI and DI, and the role of Chinese cities in investment networks. It thereby contributes to the literature tackling Chinese ‘off the map’ inter-city network research by articulating the development patterns and driving mechanisms of FDI and DI in China. Practically, the results could help urban actors to develop strategies that reflect a network vision and endogenous mechanisms. The first part of this paper introduces the development of FDI and DI in China in 2012, including size, geographical distribution, and sector composition. And it explores the relationship between cities and sectors, and compares the similarities of cities in approaching sectors. In the second part of the paper, cities and their investment are regarded as nodes and edges in networks respectively in order to carry out SNA to analyze the characteristics of investment networks. In the third part, a Negative Binomial Regression model is used to identify significant location factors that attract FDI and DI. Finally, the results are reflected upon and tentative policy recommendations are proposed.

240

14.3

14


Data and Methodology

Modelling data are derived from FDI markets,2 ORBIS,3 and China Data Online.4 There were 948 Greenfield FDI projects flowing in mainland China (excluding Hong Kong, Macao, and Taiwan) in 2012 and 17508 DI contracts recorded in ORBIS in 2012. We set the cities where source companies’ headquarters are located as source cities to simplify intricate headquarter-subsidiary structures, especially for large and fund management companies. However, by virtue of complicated ownerships and missing data, it is impossible to exclude the DI investments of firms which share ownership and are under the control of the same owners. Thus, in future empirical research it is worth delving further to understand the characteristics of investment connectivity, such as estimating the stability and structure of investment relationships, between headquarters and subsidiaries, between subsidiaries and subsidiaries, between independent firms, and between firms with shared ownership or board membership,. In the DI database, information on personal investments and the companies that have unclear headquarters and/or have ceased to operate is absent. The problem of missing addresses of source companies in the ORBIS database is dealt with by reference to information from Bloomberg and companies’ official websites. As for the location factor process, raw numbers are logged and outliers are eliminated at the outset. Afterwards, variance inflation factor (VIF) and robust standard errors are adopted to control multicollinearity and skewness respectively. The cities included are provincial-level cities (Beijing, Shanghai, Tianjin, and Chongqing), vice-province cities, and prefecture-level cities according to Chinese administrative jurisdiction. Netdraw is utilized to explore the relationship between sectors and destination cities by multidimensional scaling (MDS).5 In the SNA analysis, since the FDI network is a two-mode network, and due to a lack of outward investment data for Chinese cities, only a non-metric MDS technique is adopted to display power and similarities of nodes by means of visualization. Therefore, the focus of the networking analysis is a DI one-mode network. Firstly, the values of vertices are dichotomized and diagonal values are not considered in order to comply with principles of SNA and pay more attention to relationships rather than the strength of ties and self-investment. The SNA analysis is divided into three parts to illuminate the general pattern and individual roles in the DI network: cohesion analysis, 2 FDI Markets is a central bank of information on the globalization of business. The service tracks crossborder greenfield investment across all sectors and countries worldwide, with real-time monitoring of investment projects, capital investment and job creation. 3 ORBIS is an online database by Bureau van Dijk that contains information on over 170 million companies worldwide, with an emphasis on private company information. 4 China Data Online is an online database by China Data Center in University of Michigan that contains comprehensive statistical data of China. 5 MDS is a means of visualizing the level of similarity of individual cases in a dataset. It refers to a set of related ordination techniques used in information visualization, in particular to display the information contained in a distance matrix.

14.3

Data and Methodology

241

centrality analysis, and subgrouping analysis. Cohesion analysis includes the calculation of density, transitivity, reciprocity, geodesic distance; in centrality analysis, degree, closeness and ‘betweenness’ are adopted6; while subgrouping analysis is a bottom-up approach (cliques partition) and top-down approach (blocks and cut points, factions partition, and ‘core-periphery’ pattern).7 Since FDI and DI data are only valid for 2012, cross sectional data only are available for modelling. A negative binomial regression model is used to explore the relationship between location factors and investments since FDI and DI are over-dispersed count outcome variables. The probability mass function of the kþr 1 k negative binomial distribution is f ðk; r; pÞ PrðX ¼ kÞ ¼ p ð1 k r pÞ for k ¼ 0; 1; 2; . . . p is the probability of occurring investment; k is the number of investment projects; r is non-investment.

14.4

Results: Agglomerated Network Patterns and Significant Factors

1. FDI Development and Sector Composition Since the introduction of open policy and economic reform, massive international capital has been flowing into China. Meanwhile, foreign capital is equipped with knowledge and production modes which can be transferred to indigenous firms to improve China’s productivity and integration into global markets. Especially in the eastern areas of China, due to labour market advantages and geographic location advantages, coastal cities have formed relatively comprehensive industrial systems and account for the major share of national exports as a global manufacturing centre. Currently, China has been the largest FDI host economy in the world. This paper looks at the FDI distribution in 2012 when the world economy started to revive following the 2008 global economic crisis. Figures 14.1 and 14.2 show the top 64 FDI source cities (which invest more than 3 projects when the mean is 2.5) and the top 30 destination cities (which receive more than 5 projects when the mean is 4.6, excluding two outliers Beijing and Shanghai). It can be seen that most FDI source cities are agglomerated in North Western Europe, the US and Japan, whilst most top destination cities are agglomerated around the Chinese coastline, the 6

Degree evaluates the amount of each node’s direct linkages; closeness is each node’s sum of geodesic distances with other nodes in the network; betweenness evaluates the extent of each node locating in others’ geodesic distances. 7 Core-periphery is an ideal pattern which divides row and column into two categories. In an ideal pattern, nodes in a core block connect with each other completely so its density is 1. Nodes in a periphery block have no connections with each other so its density is 0 while they may have some connections with core nodes.

14

Fig. 14.1 Top FDI cities in the world (based on FDI markets data analysis)

242 The Conjunction of Networked Agglomeration and Location Factor …


Fig. 14.2 Top FDI destination cities in China (based on FDI markets data analysis)

14.4 243

244

14


Bohai Economic Rim, the Yangtze River Delta and the Pearl River Delta. In addition to traditional western global cities, Seoul, Singapore, and Taipei, as emerging Asian global cities, are also listed in the top 10. On the other hand, there are few star cities located in other areas, especially western China. However, some cities in the western and central areas of China are rising, such as Chengdu, Chongqing, Wuhan, Xi’an, and Changsha. Regarding FDI sector composition (Chart 14.1), there are 38 sectors that absorbed FDI in 2012. Nevertheless, an obvious gentrification pattern is displayed: the top three sectors (business service, financial services and automotive components) take account of about 30% of all investments; another 12 important sectors that absorb more than 26 investments (the mean is 25) (from industrial machinery, equipment and tools (IMET) to electronic components) take account of about 55%; 23 sectors account for the remaining investments. It is interesting to note that foreign investors put the spotlight on business services (No. 1), financial services (No. 2) and other high value-added sectors (No. 6, No. 13, and No. 15) though heavy industry sectors and manufacturing sectors still play a major role in attracting FDI. With respect to FDI activities’ structure (Chart 14.2), all FDI projects flowed into 17 activities. However, by contrast with sector composition, the structure shows considerable disparities between different activities: the top three activities (manufacturing, sales, marketing and support, and business services) represent 72% of all projects; other major activities which absorbed more than 10 projects (from retail to maintenance and services) represent around 25%; and the remaining seven activities attracted only 3%. It is therefore shown that although there is more FDI flowing into high value-added sectors, the image of a dominance of Chinese manufacturing and cheap labour remains relevant in foreign inward investment. Combining FDI sector and activity status sheds light on the following characteristics: FDI sector distribution is more scattered than activity distribution; advanced producer services (APS) sectors such as financial services, business services and other high value-added sectors including software and information technology (IT) services, and electronic components are starting to play a dominant role in attracting FDI. At the same time, secondary sector industry sectors still play a significant role, such as IMET, automotive components, transportation and chemicals; manufacturing, sales, marketing and support, and retail activities, which make up more than half of the total FDI projects. This finding indicates that labour and market size are still two crucial drivers in attracting FDI. 2. DI Development and Sector Composition Investment, consumption, and export are regarded as three carriages to drive the Chinese economy forward. In relation to investment, domestic investment makes up the major proportion and plays a significant role in upgrading and stimulating reproduction. As shown in Figure chapter 3, 60 top cities that outperform others are selected to illustrate the pattern of geographical distribution of the DI network. These cities invest more than 14 projects when the mean is 13.1, and they receive more than 48 projects when the mean is 47.8, excluding three outliers, Beijing,

Chart. 14.1 FDI sector composition (based on FDI markets data analysis)

14.4 Results: Agglomerated Network Patterns and Significant Factors 245

246

14


Chart. 14.2 FDI activity composition (based on FDI markets data analysis)

Shanghai and Shenzhen. It is shown that the majority of these top cities are located in the coastal areas, agglomerated around the Bohai Economic Rim, the Yangtze River Delta and the Pearl River Delta, similar to the FDI geography. However, besides these three city regions, the Mid-Yangtze River region and the northeast region are catching up. In addition, Chongqing and Chengdu, as two hubs in western China, are rising up to attract more DI. Besides geographical distribution, as shown in Chart 14.3, investment profile also pinpoints some characteristics of DI development in China: firstly, there is obvious gentrification in investment size between the top three cities (Beijing, Shanghai, and Shenzhen) and the others; secondly, only the top four cities (Beijing, Shanghai, Shenzhen, and Guangzhou) have an overwhelming capacity to outsource investment while the others depend heavily on inward investment (also illustrated in Fig. 14.3: the white rim indicates the scale of outsource while the red rim indicates the scale of inward investment); self-investment plays a fairly important part in many cities, especially in cities whose ranking is relatively low. The top 30 cities’ investment graph is shown in Chart 14.4, which indicates a steeper disparity between the top three cities and the others. By contrast, DI sector distribution is relatively even; the top three [Metals, Chemicals and Plastics, and Pharmaceutical and Biotechnological Products (PBP)] in 40 DI sectors make up about 24% in total (Chart 14.5); meanwhile Wholesale and Retail Trade, semiconductors and consumer electronics (SCE), and IMET also receive more than 1000 projects; the other 10 outstanding sectors that receive more


Chart. 14.3 11 DI profile of top 15 cities (based on ORBIS database analysis)

14.4 247

248

14


Fig. 14.3 Top 60 DI cities in China (based on ORBIS database analysis)

than 442 when the mean is 438, constitute around 36%. This finding denotes that natural resources, cheap labour and market size are still three dominant stimuli for DI since four out of six top sectors receive more than 1000 projects and eight out of 10 outstanding sectors strongly rely on those three factors. However, only a small number of high end sectors and APS sectors that have high added-value, lead (PBP ranks No. 4; SCE ranks No. 5; Financial service ranks No. 9; Real Estate ranks No. 13), while other value-added sectors such as Software and IT services, Aerospace, and Business Services comprise the remainder. In short, domestic investors are still mainly confined to manufacturing and heavy and light industries that are resource-intensive and labour intensive, revealing that China’s DI ability is still not sustainable even though China has become the second largest economy in the world. 3. Adjacency Between Cities and Sectors In this section, top destination cities and top sectors are selected to form a matrix so as to explore the adjacency between cities and sectors. Node size is determined by the number of projects. In order to uncover general patterns of relationships between cities and sectors, the top 50 FDI destination cities and the top six sectors (both taking over more than 50%) are dropped to form a matrix. Given the results of DI development and sector composition, the top 50 DI destination cities (more than the mean 48 projects) and the top seven sectors (both taking over more than 50%) are selected to form a matrix. As shown in Fig. 14.4, IMET and automotive components are proximate and share many destination cities. The other four sectors are relatively independent and surrounded by individual city groups. In addition, some top cities are proximate to not just one top sector, indicating that they have a more comprehensive and balanced FDI sector composition. For instance, Shanghai is


249

Chart. 14.4 Top 30 Chinese cities’ investment profile

14.4

250

14


Chart. 14.5 DI sector composition (based on ORBIS database analysis)

Fig. 14.4 FDI city-sector MDS layout (based on FDI markets data analysis)

located in the core position which is surrounded by all sectors; meanwhile, it is apparent that cities in the central area have very thick ties with sectors which express their power in this network. In addition, cities that are proximate to each other have a similar relationship with sectors. For instance, as shown in Table 14.1, major cities like Tianjin and Suzhou share a very similar investment profile, whilst

14.4


251

Table 14.1 FDI and DI MDS examples Sector city

Automotive components

Business services

Chemicals

Financial services

IMET

Software and IT services

Suzhou Tianjin Huzhou Xuzhou

6 5 0 0

2 2 0 0

4 3 1 1

3 5 0 0

17 2 1 1

1 1 1 1

Fig. 14.5 DI city-sector MDS layout (based on ORBIS data analysis)

Huzhou and Xuzhou share the same relationships with sectors. The proximity between cities illuminates the degree of similarity of profile since they are receiving investments from similar sectors. Regarding DI, MDS layout shown in Fig. 14.5, IMET, metals, and chemicals and plastics share similar destination cities while the others are relatively independent; it is found that there are no cities locating in the center of sectors any more, which indicates that most cities have DI profile preferences; in addition, powerful nodes like Beijing, Shanghai, Shenzhen, and Guangzhou are more scattered rather than clustering in the FDI-sector network, which means that big cities have more dissimilarities in the DI-sector network (Table 14.1). 4. Agglomerated Network Pattern and the Position of Cities in Networks Several SNA methods are generally carried out to unravel investment networks, mostly the DI one-mode network. For purposes of clean mapping and symmetry for the FDI network, the top 50 source cities and 50 destination cities are selected, based on the number of projects, to form a two-mode matrix; for the DI network,

252

14


Fig. 14.6 FDI city–city MDS layout (based on FDI markets data analysis)

the top 60 cities are selected to form a one-mode matrix based on the number of projects. Regarding the FDI network, due to its two-mode attribute and lack of outward FDI, only the MDS technique is used to discern the similarities of cities. As shown in Fig. 14.6, cities are clustered in the centre and are similar to each other in investment profile with Tokyo, Paris, and London as source cities for example, and Beijing and Shanghai as destination cities. In contrast, the other two Chinese mega city formations Shenzhen and Guangzhou are very different from each other. It is interesting to note here that there are some special nodes that are located far from their groups, which indicates that they have relatively unique investment preferences like Huzhou, Foshan, and Chengdu as destination cities, and Houston as a source city. In terms of DI network, city similarities evaluated by MDS are presented in Fig. 14.7 and all the results concerning its cohesion and centrality are shown in Table 14.2. It can be seen that Beijing, Shanghai, and Shenzhen are overlapped, which means that they have very similar roles in connecting other cities. In addition, cities in a central place have much denser ties with others. A good fit starts at only 60% in two dimensions, reaching 81% in nine dimensions, which means that more factors can explain the similarities of cities besides their adjacencies and closeness. Comparing the complete network, the DI network is a centralized network that has loose ties with peripheral cities in terms of degree centralization, density, and average degree. However, it is still a small world network since any city can get connected within two steps, and the longest geodesic distance is three. In terms of degree, this not only indicates the power of cities but it also highlights the different roles of cities. For instance, Guangzhou and Nanning are ‘outsiders’ as their out-degrees are much higher than their in-degrees. Chengdu is a ‘sinker’ as its in-degree is much higher than its out-degree. Ningbo is a ‘communicator’ since its

14.4


253

Fig. 14.7 DI city–city MDS layout (based on ORBIS data analysis)

in-degree equals its out-degree. In terms of closeness, Beijing, Shanghai and Shenzhen are overwhelmingly powerful in approaching all others. There are some interesting nodes that have a large gap between in-closeness and out-closeness, like Guangzhou and Nanning which can out-approach all others easily but are difficult to be approached by all others. With respect to betweenness this can explain the extent to which nodes can pass through the geodesic path between two cities. The cities with high values play a ‘broker’ role in this network. Among four cities, Guangzhou’s value is relatively low, which indicates that its broker power is limited in a subgroup rather than the overall network like Beijing, Shanghai, and Shenzhen. It is surprising that Jinan performs much better than the same level of cities, illustrating its powerful broker role in the network. In addition, Nanning is special again in terms of its poor ‘bridging’ capacity. Transitivity is expressed by a clustering coefficient which calculates the density of its open neighborhood. Such cities have low values since their linked cities have less dense ties with each other. The cities with high value like Shaoxing, Dalian, Suzhou and Changsha have ego networks that are more cohesive and have high reachability. In terms of asymmetric value of overall network and individuals, the DI network is a reciprocate network, particularly in big cities. Besides cohesion and centrality analysis, this paper also adopts several subgrouping methods to investigate whether there are some factions in the DI network. Firstly, clique8 partition, as a bottom-up approach, is carried out. It is interesting to look at co-membership across cliques which can denote the capacity of cities in

8

A clique is a sub-set of a network in which the actors are more closely and intensely tied to one another than they are to other members of the network.

In-degree

Out-degree

In-closeness

Out-closeness

Betweenness

Beijing 41 58 75 58 653.348 Shanghai 39 55 77 61 501.621 Shenzhen 30 57 87 59 336.1 Guangzhou 19 53 100 63 132.086 Tianjin 20 35 98 81 131.653 Hangzhou 26 27 91 89 97.975 Chongqing 17 32 102 84 83.293 Nanjing 23 27 95 89 61.336 Ningbo 17 17 100 99 33.318 Changsha 15 19 104 97 19.319 Nanning 5 28 116 88 2.8 Fuzhou 15 23 105 93 39.166 Suzhou 21 16 97 100 38.367 Wuhan 23 15 93 101 55.878 Jinan 14 20 105 96 94.694 Chengdu 24 12 93 104 26.186 Dalian 13 16 105 100 18.14 Nanchang 15 15 103 101 40.171 Shaoxing 18 9 98 107 11.23 Changzhou 15 16 102 100 28.67 Density = 0.247, Av-degree = 14.322, degree centralization = 0.779, Av-distance = 1.807, diameter = 3, asymmetric

Top 20 cities

Table 14.2 Cohesion and centrality of cities in the DI network (based on ORBIS data analysis) 0.226 0.239 0.235 0.254 0.324 0.337 0.324 0.366 0.442 0.454 0.393 0.375 0.472 0.376 0.354 0.403 0.496 0.413 0.494 0.451 = 0.274

Transitivity 0.707 0.679 0.526 0.309 0.41 0.359 0.256 0.389 0.308 0.214 0.138 0.267 0.48 0.226 0.214 0.2 0.318 0.364 0.174 0.292

Reciprocity

254 14 The Conjunction of Networked Agglomeration and Location Factor …

14.4


255

bridging cliques. When clique size is set at five, 32 cliques are found. The rank of co-membership is as follows: Beijing (32), Shanghai (29), Shenzhen (16), Nanjing (11), Guangzhou (8), Tianjin (8), Hangzhou (7), Chongqing (3), Changsha (3), and Xi’an (2). When clique size is set at six, eight cliques are found. The rank of co-membership is as follows: Beijing (8), Shanghai (8), Shenzhen (8), Guangzhou (6), Nanjing (3), Hangzhou (3), Tianjin (3), and Chongqing (2). Secondly, through block and cut-point9 analysis, no blocks and cut-points are found in the DI network. Lastly, this paper carries out factions partitions clustering analysis through Tabu optimization and core-periphery pattern in order to find closely connected subgroups in the DI network. Through iterating algorithms in each analysis, they have a similar result, indicating that instead of several cohesive subgroups existing in the DI network, there is one major relatively firm and cohesive subgroup (of core players). Figure 14.8 illustrates the combination of core-periphery pattern (fitness = 0.578) and 10 factions when poor fit and faction number are relatively low, iterating from five factions to 18 factions. The factions are distinguished in 10 colours; blue nodes are core players while other colours are peripheral players; the only difference is that Jinan, Wuhan and Nanchang are core players in core-periphery analysis but are dropped in the dark blue faction, whilst Xuzhou is a peripheral player but is dropped in the light blue faction (core faction). When geographical proximity is considered, the nodes that exist in the same factions in a network sense and are close to each other in a geographical sense can be explained as geographical factions or cohesive city regions: such as Huizhou and Zhongshan located in both the yellow faction and the Pearl River Delta, Putian and Quanzhou located in both the purple faction and Fujian Province, Tangshan and Handan located in both the brown faction and Hebei Province, Foshan and Zhuhai located in both the orange faction and the Pearl River Delta, Haerbin, Dalian and Shenyang located in both the dark purple faction and northeast region, Jinan, Qingdao and Zibo located in both the dark blue faction and Shandong Province, Shanghai, Hangzhou, Nanjing, Suzhou, Ningbo, Shaoxing, Changzhou, Wuxi and Xuzhou located in both the core faction and the Yangtze River Delta, Bejing and Tianjin located in both the core faction and the Bohai Rim Economic Zone, and Shenzhen and Guangzhou located in both the core faction and Pearl River Delta. Therefore, among the regions, the Yangtze River Delta is a more cohesive and balanced region since it has more core players agglomerated in its region while the other two are more loose and centralized urban regions. 5. Identifying Significant Factors Besides SNA methods that analyze the relationships and roles of cities in networks, location factor analysis is the other fundamental dimension that can explain agglomerating FDI and DI in different cities. Next, Negative Binomial Model is considered to explore the relationship between location factors and investments. As

9

Cutpoints are the nodes that divide a graph into two disconnected sub-graphs. The divisions into which cutpoints divide a graph are called blocks.

14

Fig. 14.8 Faction partitions in the top 60 cities (based on ORBIS data analysis)


14.4


257

shown in Table 14.3, 14 models have been used to identify factors significant to FDI. In M1, DI is identified as very significant to FDI when the model fit is 16.2%. In M3, population growth rate is negatively related to FDI while population density is positively related to FDI when the model fit is 13.1%. In M4, the average number and wages of employees are both very significant to FDI when the model fit is 20.7%. In M5, we can see that only employees in secondary sector industries, as opposed to primary and tertiary sector industries, is very significant to FDI when the model fit is 19.3%. In M6, banking and insurance is significant to FDI instead of (anticipated) scientific research. In M7, fixed assets investment is detected as a very significant contributor to FDI when the model fit is 24.7%. In M8, gross industrial output of foreign funded enterprises is identified as a very significant factor for FDI when the model fit is 14.4%. In M9, students enrolled in higher education institutions is very significant to FDI when the model fit is 11.4%. In M10, the area of paved roads is identified as a very significant factor to FDI when the model fit is 23%, meanwhile electricity consumption is also considered a significant factor to FDI regardless of its significance level of 5% since it is also identified in M12, M13, and M14. In M11, dummy variable ‘inland/coast’ location is very significant to FDI when the model fit is 3.8%. In M12, the average wages of employees (significance level of 1%), employees of secondary sector industry (significance level of 1%), gross industrial output of foreign funded enterprises (significance level of 5%), students enrolled in higher education institutions (significance level of 5%), and electricity consumption (significance level of 5%) are identified again in different significant levels when the model fit is 27.8%. In M14 with the dummy variable inland/coast, all the significant factors identified in M13 with the same significance level plus inland/coast (significance level of 5%), are significant to FDI when the model fit is 28.5%. In the overall final model with fixed effect of Province, gross industrial output of foreign enterprises and students enrolled in higher education institutions are identified as two very significant factors to FDI again, whilst population density and electricity consumption are identified at a 5% significance level. Based on the results of these FDI models, it can be deduced that highly skilled labour, foreign capital accumulation (path dependence), and market size are three key drivers in FDI agglomeration. It is also found that DI has a crowd-in effect on FDI. Meanwhile, the strength of secondary sector industry, including manufacturing, light industry, and heavy industry, is still a basic condition attracting FDI, which may be explained by the fact that the majority of FDI still flows into manufacturing-related sectors that are labour intensive and resource intensive. It is interesting that financial services have become a significant factor in the agglomeration of FDI projects, suggesting that the capital source of FDI projects tends to be localized in supply. In addition, the size of fixed assets investment and paved roads not only indicates the dynamics of local economies but also demonstrates the local capacity of reproduction and developed infrastructure to some extent. Therefore, the dynamics of the local economy and a well-developed industrial base and urban infrastructure are critical to agglomerate FDI. With respect to geographic

0.0112***

DI

Scientific research

Banking and insurance

Electric power gas water

Employ. tertia

Emply. secon

Employ. prim

Average wage staff workers

Av. Employees

Population density

(0.84)

0.00120

(−1.10)

−0.426

FDI

(M2) FDI

(M7) FDI

(M8) FDI

(M9) FDI

(M10) FDI

(M11) FDI

(M12)

(4.07)

(6.35)

(−0.03)

0.339

(−0.90)

−0.150

(3.27)

0.950**

(−0.60)

(3.13)

(8.04) −0.000128

0.0190**

0.0529***

(−1.49)

0.0000897***

0.000118***

(4.16)

(1.62)

−0.192

FDI

(M6)

(7.48)

−0.123

FDI

(M5)

0.000554

0.0202***

FDI

(M4)

0.00374***

(−3.04)

−0.0937**

FDI

(M3)

(2.71)

0.0175**

(3.70)

(−0.47)

−0.0141

(0.54)

0.0000232

(2.50)

0.00139*

FDI

(M14)

(continued)

0.0000758***

(1.39)

0.000422

FDI

(M13)

14

Population growth rate

Developed land area

Land area

FDI

Dependent variable

(5.30)

(M1)

Model

Table 14.3 FDI model results


185

234

(7.39)

1.138***

(−3.28)

−1.112**

234

(2.31)

0.470*

(−8.50)

−3.863***

0.207

Negative Binomial

(M4)

233

(2.07)

0.476*

(−5.54)

−1.125***

0.193

Negative Binomial

(M5)

234

(4.91)

0.904***

(−5.13)

−1.383***

0.161

Negative Binomial

(0.92)

(M6)

(M11)

(M12)

0.994***

224

230

(4.81)

−0.167 (−0.75)

(−2.32)

−0.448*

0.144

Negative Binomial

230

(6.55)

1.265***

(−2.70)

−0.700**

0.114

Negative Binomial

(0.98)

231

(0.97)

0.251

(−7.01)

−1.763***

0.23

Negative Binomial

(0.68)

234

(11.66)

1.827***

(0.24)

0.0762

0.038

221

(−1.45)

−0.379

(−6.58)

−3.760***

0.278

221

(−1.82)

−0.421

(−6.59)

−3.620***

0.285

Negative Binomial

(2.32) Negative Binomial

0.533* (4.01)

(2.23)

0.000635*

(0.78)

0.0000743

(2.53)

0.0233*

(2.24)

0.000206*

(−2.17)

−0.000903*

(M13)

2.043***

Negative Binomial

(2.24)

(2.44) 0.000142

0.000715*

(0.62) 0.00167*

0.0000615

0.000618*** (8.47)

(2.25)

(5.25) 0.00884

0.0170*

(2.36) 0.0874***

0.000227*

0.00135*** (4.53)

(0.24)

(−8.09)

−1.345***

0.247

Negative Binomial

(0.87)

0.0000695

−0.000976

(M10)

(−1.79)

(M9)

0.000759*

0.0000718

(M8)

(2.18)

(7.01)

0.00143***

(M7)

221

(−2.92)

−3.768**

0.99

Province Fixed

(−2.25)

−0.00273*

(0.57)

0.000242

(3.53)

0.101***

(5.32)

0.00554***

(1.04)

0.00173

(M14)

Encode average number of employees (Av. Employees), employees primary industry (Employ. prim), employees secondary industry (Emply. secon), employees tertiary industry (Employ. tertia), newly signed contracts by foreign capital (Sigcontr), gross ind. output HK funded firms (GIO. HK), Gross ind. output foreign funded firms (GIO. Foreign), student enrolment higher education (Student. HE), student enrolment secondary education (Student. SE), electricity consumption Per capita (Elec. consump), post and telecommunication offices (Post. tele), inland or coastal cities, coastal = 1, (Indcoast)

234

2.085***

(11.83)

0.906***

(3.81)

(−4.06)

(4.46)

1.712***

0.131

Negative Binomial

(M3)


N

Inalpha constant

Constant

0.006

Negative Binomial

(M2)

−0.828***

0.162

McFadden R2

R2

Negative Binomial

(M1)

Model Specification

Indcoast (dummy)

Post.tele

Elec. consump

Area paved roads

Student. SE

Student. HE

GIO. Foreign

GIO. HK

Loans

Sigcontr

Fixed assets investment

Environmental management

Model

Table 14.3 (continued) 14.4 259

260

14


location considered with reference to the location of FDI projects, a coastal advantage is still evident. Table 14.4 illustrates DI significant location factors by means of the same method. The results show that the DI and FDI models share some significant factors, including population density, average number and wages of employees, employees in secondary sector industry, banking and insurance, fixed assets investment, students enrolled in higher education institutions, and the area of paved roads, on the same significance level but with less model fit. In contrast to FDI heavy reliance on DI, DI is less reliant on FDI at a 5% significance level. Meanwhile in M2, it is interesting that land area is identified as a very significant factor, but negatively. In addition, a really surprising result is that employees in primary industry, is seen as a very significant factor negatively in M5, M12, and M13. Like gross output of foreign enterprises’ contribution to FDI, gross output of domestic enterprises attracts DI at a very significant level. In addition, some factors that were not identified in the FDI models are recognized as significant factors in the DI models, including loans (on a 1% significance level) and gross industrial output of HK enterprises (on a 5% significance level). In addition, electricity consumption is identified as a more significant factor in M12 and M13 when the model fits are 11.6 and 11.7% respectively, compared to its significance in the FDI models. With respect to lower model fits, the significant factors for DI are more intricate and easily affected by non-economic conditions. Generally speaking, it is concluded that labour force, market size, industrial base and capital accumulation, and institutional context, are significant to agglomerate DI. Significant banking and loans factors not only demonstrate the critical role of financial services in agglomerating DI but also indicate the importance of institutional context due to the regulatory control over banking and the financial system of governments. In addition, it is shown that industrial structure is more significant for DI than for FDI, because apart from the significance of secondary sector industry, agriculture plays a very significant negative role in DI. Combining significant land area may indicate that the development of China is still affected by a shortage of urban land supply or urban expansion, and that the dominance of traditional agriculture is a drag on the urbanization of China that is needed to facilitate a transformation to modern agriculture and liberate an increased labour supply to cities. Another interesting finding is that Hong Kong’s closer relationship with DI instead of FDI, suggests that Hong Kong’s capital has a greater agglomeration effect on DI.

14.5

Conclusions and Recommendations

During economic transition, China is confronted with the danger of a hard landing due to its shrinking FDI and present global economic malaise. China has slowed down its astonishing growth and begun to take steps to liberalize interest rates and reduce its dependency on exports. Since cities are equipped with abundant human

0.0669*

FDI

(0.24)

0.0000818

(−4.46)

−0.138***

DI

(M2) DI

(M7) DI

(M8)

(M9) DI

(M10) DI

(M11) DI

(M12) DI

(0.39)

(−0.60)

(3.76)

0.388

(−0.47)

−0.0709

−0.0213

0.462***

(0.44)

(2.07)

0.00260

0.00990*

(6.02)

(−4.65)

(−7.03) 0.0393***

−0.123***

(1.53)

(5.84) −0.157***

0.0000231

0.0000819***

(4.08)

(0.07)

0.0817

DI

(M6)

(5.70)

DI

(M5)

0.0000126

0.0129***

DI

(M4)

0.00221***

(−1.75)

−0.0459

DI

(M3)

(M13)

(−0.51)

−0.0192

(2.13)

0.00970*

(−4.60)

(M14)

(−0.52)

−0.431

DI

(continued)

−0.121***

(1.66)

0.0000249

(0.45)

0.0000864

DI


Environmental management

Scientific research

Banking and insurance

Electric power gas water

Employ. tertia

Emply. secon

Employ. Prim

Average wage staff workers

Av. Employees

Population density

Population growth rate

Developed land area

(2.21)

DI

Dependent variable

Land area

(M1)

Model

Table 14.4 DI model results 14.5 261

(6.47)

185

(4.16)

234

0.760***

0.397***

234

(5.17)

0.414***

(11.55)

234

(1.10)

0.0893

(3.65)

Encode gross industrial output of domestic funded firms (GIO. domestic)

N

Inalpha constant

(17.22)

(25.48)

1.149***

0.081

233

234

(1.73)

0.151

−0.0482 (−0.54)

(17.27)

2.665***

0.074

Negative Binomial

(1.43)

(M6)

(25.77)

2.853***

0.096

Negative Binomial

(M5)

224

(−0.61)

−0.0590

(25.71)

2.851***

0.094

Negative Binomial

(2.87)

0.000272**

(4.04)

0.000691***

(M7)

(M12)

230

(1.30)

0.105

(25.78)

2.850***

0.078

Negative Binomial

230

(1.91)

0.284

(18.97)

3.198***

0.057

Negative Binomial

(0.55)

231

(−0.33)

−0.0309

(16.26)

2.509***

0.094

Negative Binomial

(−0.05)

−0.0000104

234

(7.75)

0.633***

(27.32)

3.680***

0.015

227

(−2.79)

−0.271**

(5.88)

2.113***

0.116

(−1.31)

227

(−2.84)

−0.280**

(5.85)

2.077***

0.117

Negative Binomial

(4.02) Negative Binomial

−0.213

(3.13)

0.000875**

(−0.23)

−0.0000219

(4.09)

0.0290***

(2.96)

0.000203**

(M13)

1.121***

Negative Binomial

(3.48)

(2.35)

(−0.04) 0.000792***

(7.90) 0.00117*

−0.00000407

0.000470***

(4.34)

(7.07) 0.00352

0.0306***

0.0617***

(3.25)

(M11)

0.000168**

(M10)

(9.58)

(M9)

0.000557***

(2.45)

0.000271*

(M8)

Province Fixed

(1.86)

0.208

(3.38)

3.162***

(M14)

229

(−1.06)

−32.28

4.510***

3.610***

3.122***

0.043

Negative Binomial

(M4)

0.889

0.033

Negative Binomial

(M3)

Constant

0.45

McFadden R2

Negative Binomial

(M2)

R2

Negative Binomial

Model Specification

Indcoast (dummy)

Post. tele

Elec. consump

Area paved roads

Student. SE

Student. HE

(M1)

14

GIO. domestic

GIO. HK

Loans

Fixed assets investment

Model

Table 14.4 (continued)


14.5


263

capital, financial capital, instructional capital and social capital, they are the main areas where the nation’s industrial base is upgraded and where the circulation of resources is happening. Both in FDI and DI networks, outperforming cities display an agglomerated pattern in a geographical sense, just as top foreign cities show coastal metropolitan agglomeration in the US, North Western Europe, and the West Pacific, China has been experiencing high domestic agglomeration in coastal areas. 1. Upgrade the Value Chain and Keep Eyes on Significant Location Factors According to FDI sector composition, FDI projects are highly clustered in tertiary APS sectors although secondary sector industries still have an important role. However, APS and other sectors look to highly skilled labour and market size as key investment factors. In DI sectors, manufacturing and heavy and light industries that are traditionally dependent on natural resources, cheap labour and market size, remain dominant. In terms of the relationship between cities and sectors, FDI-large cities are more centralized and balanced while DI-large cities have more preferences in investment profile. In short, low valued-added activities and labour intensive sectors are still mainstream in agglomerating Chinese investment. Evidence from FDI Markets suggests that nowadays, some international investors are retreating from China’s markets since its low-cost labour advantage is weakening. In consequence, if China wishes to maintain its global position in attracting FDI, it has to expand and develop its knowledge-based sectors such as APS and high end activities that create major added value and stimulate industrial innovation. Hence, local social knowledge management and the development of a regional mindset that encourages knowledge sharing and collaborative learning are likely to be of key importance in stimulating regional and national development (see Asheim and Isaksen 2002; Gertler and Wolfe 2004). In addition to the role of APS and high end activities in upgrading value chains, several other significant factors should be considered in FDI and DI agglomeration. Regarding the model results, industrial base and infrastructure, capital accumulation, market size, and financial support are identified generally as significant location factors agglomerating FDI and DI. Therefore, cities are encouraged to improve their reproduction capacity and to cooperate with scientific institutions, for example by establishing industrial parks and research centres. Given the effect of capital accumulation, local governments also need to put forward incentivization policies for foreign investors such as the promotion of enhanced business atmosphere, land policy, tax policy and infrastructure. Thirdly, local governments should reform the Household Registration System (HRS) and offer fair social services to city outsiders to improve their sense of belonging and so facilitate labour mobility and expand the market size. Meanwhile, the education system should be diversified. For instance, local governments can encourage the building of private schools, higher education and training establishments that would promote a qualified labour force. Lastly, financial services are more and more important in driving the local economy, so China should loosen the control of the financial system and improve its openness in order to promote specialized services, especially for medium and

264

14


small enterprises. In addition, agricultural modernization that could be facilitated by the agglomeration of advanced services to keep the Chinese economy healthy and sustainable cannot be ignored. 2. A Need to Recognize City Positions and Make Networking Strategies Recognition of Chinese city positions and the identification of sub-groups in networks are critical to efficient actions to improve opportunities and overcome constraints. In terms of similarities of investment profile, city administrations need to pay attention to other cities which share a similar investment profile because they are striving for investment from the same sources. In order to improve China’s social and economic sustainability, cities can establish alliances with complementary partners that can integrate resources and circulate information efficiently. The cultivation of trust and creativity are essential preconditions for successful city alliances. Regarding the centrality analysis, city nodes play different roles in the network such as sinker, outsider, communicator, and broker. There are no absolute advantages or disadvantages implicit in these diverse roles. Nevertheless, the nodes with high degrees are more inter-linked with other highly inter-linked cities. Meanwhile, brokerage is regarded as one kind of social capital because the broker role establishes relationships between two groups that have heterogeneity and opportunity and it provides access to valuable information and translation fees by bridging isolated nodes (Burt 2009). Therefore, cities that are extremely outbalanced and poor in bridging others tend to fall apart and lose network connectivity. They should build more effective ties with significant nodes and cultivate a capacity to build bridges for others. Some scholars criticize the overemphasis on structural holes. For instance, in technical collaboration networks, increasing structural holes are contended to reduce innovation output (Ahuja 2000). The optimal structure of networks is contingent on the objectives and content of relationships which still need further empirical research. But if core cities are to maintain their core roles in the long run, they must retain their attractiveness to elites and promote cooperation with other core cities so as to consolidate their network position. Meanwhile, they also need to develop their periphery branches and consolidate their role in ego networks. It is argued that shared information is happening indigenously whilst membership and association fees are applied to keep privilege in the local network (Carroll 2007). The spill-over effect is not obvious in high-end industries whilst information circulation happens indigenously and spontaneously (Nicolini 2003). Hence, in terms of cities in the periphery, they should build more linkages with cities in the core to get involved in the highly connected group of cities. In addition, they should cooperate with each other so as to form a new cohesive city network group. Focusing on city network functional complementarities at a city region level would stimulate Taylor and Pain’s Jacobsean economic expansion Process A whereby smaller towns and cities surrounding significant networked nodes for inward investment are enveloped and upgraded in highly internally and externally interconnected global mega-city regions.

References

265

References Agosin, M.R., and R. Machado. 2007. Openness and the international allocation of foreign direct investment. The Journal of Development Studies 43 (7): 1234–1247. Ahuja, G. 2000. Collaboration networks, structural holes, and innovation: A longitudinal study. Administrative Science Quarterly 45 (3): 425–455. Alfaro, L., A. Chandab, and S. Kalemli-Ozcan. 2004. FDI and economic growth: The role of local financial markets. Journal of International Economics 64 (1): 89–112. Anderson, D. 1990. Investment and economic growth. World Development 18 (8): 1057–1079. Asheim, B.T., and A. Isaksen. 2002. Regional innovation systems: the integration of local ‘sticky’ and global ‘ubiquitous’ knowledge. The Journal of Technology Transfer 27 (1): 77–86. Barrell, R., and N. Pain. 1997. Foreign direct investment, technological change, and economic growth within Europe. The Economic Journal 107 (445): 1770–1786. Barro, R.J. 1989. Economic growth in a cross section of countries. National Bureau of Economic Research 3120. Berry, B.J. 1964. Cities as systems within systems of cities. Papers in Regional Science 13 (1): 147–163. Blomstrom, M., Globerman, S. and A. Kokko. 1999. The determinants of host country spillovers from foreign direct investment: Review and synthesis of the literature. No. 502 SSE/EFI Working Paper Series in Economics and Finance, Stockholm School of Economics. Blomström, M., and F. Sjöholm. 1999. Technology transfer and spillovers: Does local participation with multinationals matter? European Economic Review 43 (4): 915–923. Borensztein, E., J. De Gregorio, and J.W. Lee. 1998. How does foreign direct investment affect economic growth? Journal of International Economics 45 (1): 115–135. Borgatti, S.P., and P.C. Foster. 2003. The network paradigm in organizational research: A review and typology. Journal of Management 29 (6): 991–1013. Boschma, R. 2004. Competitiveness of regions from an evolutionary perspective. Regional Studies 38 (9): 1001–1014. Burt, R.S. 2009. Structural holes: The social structure of competition. Cambridge: Harvard University Press. Cairncross, F. 2001. The death of distance: How the communications revolution is changing our lives. Boston: Harvard Business Press. Capello, R. 2000. The city network paradigm: Measuring urban network externalities. Urban Studies 37 (11): 1925–1945. Carroll, W.K. 2007. Global cities in the global corporate network. Environment and Planning A 39 (10): 2297. Castells, M. 1996. The rise of the network society: The information age: Economy, society and culture, vol. I. Malden: Blackwell. Caves, R.E. 1971. International corporations: The industrial economics of foreign investment. Economica 38: 1–27. De Mello, L.R. 1999. Foreign direct investment-led growth: Evidence from time series and panel data. Oxford Economic Papers 51 (1): 133–151. Deok-Ki Kim, D., and J.-S. Seo. 2003. Does FDI inflow crowd out domestic investment in Korea? Journal of Economic Studies 30 (6): 605–622. Domar, E.D. 1946. Capital expansion, rate of growth, and employment. Econometrica, Journal of the Econometric Society: 137–147. Dunning, J.H. 1981. Explaining the international direct investment position of countries: Towards a dynamic or developmental approach. Weltwirtschaftliches Archiv 117 (1): 30–64. Ellison, G., and E.L. Glaeser. 1999. The geographic concentration of industry: Does natural advantage explain agglomeration? American Economic Review: 311–316. Fingleton, B. 1999. Estimates of time to economic convergence: An analysis of regions of the European Union. International Regional Science Review 22 (1): 5–34.

266

14


Florida, R. 2002. The economic geography of talent. Annals of the Association of American Geographers 92 (4): 743–755. Glaeser, E.L. 2010. Agglomeration economics. Chicago: University of Chicago Press. Glass, A.J., and K. Saggi. 1998. International technology transfer and the technology gap. Journal of Development Economics 55 (2): 369–398. Gottman, J. 1961. Megalopolis: The urbanization of the Northeastern Seaboard of the United States. Cambridge: The MIT Press. Håkansson, H., and I. Snehota. 1989. No business is an island: The network concept of business strategy. Scandinavian Journal of Management 5 (3): 187–200. Hall, P., and K. Pain (eds.). 2006. The polycentric metropolis: Learning from mega-city regions in Europe. London: Earthscan. Helsley, R.W., and W.C. Strange. 1990. Matching and agglomeration economies in a system of cities. Regional Science and Urban Economics 20 (2): 189–212. Henderson, J.V. 1991. Urban development: Theory, fact, and illusion. Oxford: Oxford University Press. https://ideas.repec.org/b/oxp/obooks/9780195069020.html. Henderson, J.V. 2003. Marshall’s scale economies. Journal of Urban Economics 53 (1): 1–28. Isard, W. 1956. Location and space-economy: A general theory relating to industrial location, market areas, land use, trade, and urban structure. The Technology Press of Massachusetts Institute of Technology and Wiley.http://trid.trb.org/view.aspx?id=131509. Jacobs, J. 1969. The economy of cities. New York: Random House. Jacobs, J. 1984. Cities and the wealth of nations. New York: Random House. Jacobs, W., H. Koster, and P. Hall. 2011. The location and global network structure of maritime advanced producer services. Urban Studies 48 (13): 2749–2769. Johansson, B. and J.M. Quigley. 2004. Agglomeration and networks in spatial economies. In Fifty years of regional science, ed. Florax, R. and Plane, D.A., 165–176. Berlin: Springer. Kim, S. 1995. Expansion of markets and the geographic distribution of economic activities: The trends in US regional manufacturing structure, 1860–1987. The Quarterly Journal of Economics: 881–908. Krugman, P.R. 1997. Development, geography, and economic theory. Cambridge: The MIT Press. Liu, Z. 2008. Foreign direct investment and technology spillovers: Theory and evidence. Journal of Development Economics 85 (1): 176–193. Mahroum, S., R. Huggins, N. Clayton, K. Pain, and P.J. Taylor. 2008. Innovation by adoption: Measuring and mapping absorptive capacity in UK nations and regions. National Endowment for Science, Technology and the Arts (NESTA). http://www.nesta.org.uk/assets/Uploads/pdf/ ResearchReport/innovation_by_adoption_report_NESTA.pdf. Marshall, A. 1920. Principles of economics: An introductory volume. Macmillan and Co. Ltd. http://www.econlib.org/library/Marshall/marP.html. Nicolini, D., S. Gherardi, and D. Yanow. 2003. Knowing in organizations: A practice-based approach. Armonk: ME Sharpe. Noorzoy, M. 1979. Flows of direct investment and their effects on investment in Canada. Economics Letters 2 (3): 257–261. O’Brien, R. 1992. Global financial integration: The end of geography. London: Royal Institute of International Affairs. Ouyang, P., and S. Fu. 2012. Economic growth, local industrial development and inter-regional spillovers from foreign direct investment: Evidence from China. China Economic Review 23 (2): 445–460. Pain, K. 2008a. Examining core-periphery relationships in a global mega-city region—The case of London and South East England. Regional Studies 42 (8): 1161–1172. Pain, K. 2008b. Spaces of practice in advanced business services: Rethinking London-Frankfurt relations. Environment and Planning D Society and Space 26 (2): 264. Pain, K., and G. Van Hamme (eds.). 2014. Changing urban and regional relations in a globalizing world: Europe as a global macro-region. Cheltenham: Edward Elgar. Porter, M.E. 1990. The competitive advantage of nations. Harvard Business Review 68 (2): 73–93.

References

267

Powell, W.W., K.W. Koput, and L. Smith-Doerr. 1996. Interorganizational collaboration and the locus of innovation: Networks of learning in biotechnology. Administrative Science Quarterly: 116–145. Romer, P.M. 1986. Increasing returns and long-run growth. The Journal of Political Economy: 1002–1037. Scott, A. (ed.). 2001. Global city regions. Oxford: Oxford University Press. Tang, S., E.A. Selvanathan, and S. Selvanathan. 2008. Foreign direct investment, domestic investment and economic growth in China: A time series analysis. The World Economy 31 (10): 1292–1309. Taylor, P.J. 2012. Extraordinary cities: Early ‘city-ness’ and the origins of agriculture and states. International Journal of Urban and Regional Research 36 (3): 415–447. Taylor, P.J., and K. Pain. 2007. Polycentric mega-city regions: Exploratory research from Western Europe. In The Healdsburg research seminar on megaregions, ed. Todorovich, P., 59–67. Lincoln Institute of Land Policy and Regional Plan Association. http://library.rpa.org/pdf/ 2050-The-Healdsburg-Research-Seminar-on-Megaregions-2007.pdf. Turok, I. 2004. Cities, regions and competitiveness. Regional Studies 38 (9): 1069–1083. United Nations. 2013. World population prospects: The 2012 revision. Population Division New York, Department of Economic and Social Affairs, United Nations. Van Oort, F.G. 2007. Spatial and sectoral composition effects of agglomeration economies in the Netherlands. Papers in Regional Science 86 (1): 5–30. Wall, R.S., and G. Van der Knaap. 2011. Sectoral differentiation and network structure within contemporary worldwide corporate networks. Economic Geography 87 (3): 267–308. Wolfe, D.A., and M.S. Gertler. 2004. Clusters from the inside and out: Local dynamics and global linkages. Urban Studies 41 (5–6): 1071–1093.